By Daniel Miller, Chief Credit Officer
Summary
During periods of elevated default rates, decreases in leveraged loan prices play an upsized role as an important driver of CLO equity returns.
Depressed loan prices allow CLOs to engage in “par-building:” the creation of additional cash flow to the CLO equity holder by reinvesting incoming cash from loan maturities and prepayments in loans with temporarily depressed prices.
The relationship between loan prices and default rates during the past 5 default cycles provides significant insight into par-building opportunities for the 2023-2024 credit default cycle.
Based on Capra’s 3.8% default forecast and 93.5 average loan price forecast, par-building will significantly mitigate credit losses during the upcoming credit default cycle which we expect to last through 2024.
CLO Features Enable Mitigation of Loan Defaults
As default rates increase, with downgrades outweighing upgrades by a 3:1 rate, most credit forecasters are predicting a default cycle with default rates surpassing long-term averages this year and into 2024. In January, Capra’s leveraged loan default forecast called for a default rate of 3.8% in 2023 and 3.3% in 2024 – both well above the long-term average [1].
While CLOs and especially equity tranche investors suffer during periods of above average defaults, they can offset those losses with a structural safeguard specific to CLOs called “par-building.” Par-building refers to the ability of CLO managers to invest both scheduled and early principal loan payments into newly purchased loans. In periods of higher loan defaults, high-quality loans can be available at prices well below par. When these loans ultimately rise in price, the excess cash flows through to the CLO equity holder, potentially offsetting losses from loan defaults.
Yet, Not All Default Cycles Offer Like Opportunities to Par-Build
Par-building requires a drop in average loan prices well below the historic average.
In Graph 1, we look at the relationship between loan price (x-axis) and default rate (y-axis). The blue dots on the graph show the average loan price and average default rate during each of the default cycles of 2002, 2008, 2016, 2019, 2020 and Capra’s 2023-2024 projection. The “Non-Default Cycle Average” dot denotes the average loan price and average default rate for all the periods outside of the default cycles. The bent line denotes the points at which the positive impact of par-building offsets the losses from credit defaults [2].
The default cycles denoted by the blue dots above the line (the red area) had limited par-building opportunity in relation to the default magnitude. The default cycles below the line (the green area) offered ample par-building when compared to the magnitude of the default rate.
While the default cycles of 2002 and 2020 were not conducive to par-building in relation to the default rate, the 2008 cycle was very conducive to par-building. The 2019 default cycle was not conducive to par-building while the 2016 default cycle offered modest par-building opportunity even though it was a very short cycle.
The 2023-2024 anticipated default cycle uses Capra Ibex’s 2023-2024 default forecast as well as our internal forecast for 93.5 average loan price during this default cycle. These forecasts indicate that parbuilding will be enough to offset losses from defaults for CLO deals with ample remaining re-investment periods.
Graph 1
Better Understanding Previous Default Cycles
To put the default cycles into context, Table 1 characterizes the attributes that drove the spike in default rate for each cycle.
Table 1
Table 2 shows the timeline of each default cycle and the data behind the blue dots in Graph 1, including average loan price and the average default rate. It also shows the average loan price and default rate during the periods outside of the default cycles.
Table 2 [3]
Chronological Review of Previous Default Cycles
2000 - 2003 Default Cycle
The 2000-2003 default cycle was unique in its longevity, resulting in a leveraged loan default rate of 7.5% in 2000, 7.1% in 2001, 5.9% in 2002, and an annualized default rate of 5.6% for the first 5 months of 2003. During this time, loan prices were surprisingly resilient, averaging 95.0 (less than 1% below the long-term average of 95.6 and 2.4% below the non-default cycle average).
A mild recession turned into the worst cumulative default cycle of the 21st century because of overinvestment in telecommunications infrastructure (including long-haul fiber-optic networks) coupled with a series of prominent accounting scandals resulting in defaults. Additionally, there was a brief recession following a long period of expansion in the 1990s, exacerbated by the bursting of the dot.com bubble and the September 11 terrorist attack. Since the causes of the default rate were disparate and narrower in scope, average loan prices stayed close to the long-term average.
Had CLOs been prevalent, this would have been a terrible default cycle for par-building. In addition, the duration of the cycle would have exacerbated the negative impact for CLO equity investors.
2008 - 2010 Default Cycle
The Great Financial Crisis and subsequent recession ran from September 2008 until September 2010. The average loan price of 82.1 was well below the long-term non-default cycle average of 97.4, allowing for plenty of par-building. The default rate was 10.3% the first year and 4.8% the second year. The default cycle was triggered by a near collapse of the financial system followed by a broad and deep recession across most industries.
The 2008-10 default cycle ranks as the best performing when weighing the negative impact of the default rate against the positive impact of par-building.
2016 Default Cycle
The 2016 default cycle was a short one and driven by a collapse in crude prices resulting in defaults in the oil and gas sectors. Intex data for CLO loan prices (which first became available around this time) in Table 3 show how concentrated the default cycle was in the energy sector (and in the mining sector, which was only 1% of the portfolio) [4]. Even with the concentrated nature of defaults within one industry, the price declines were severe enough to bring down the average loan price and create a moderate par-building opportunity.
Table 3
2019 Default Cycle
The 2019 default cycle was not really a default cycle by traditional measures and its duration was only a few months. However, the combination of the default rate exceeding the historic average default rate and the resiliency of the average loan prices negatively impacted CLO equity tranches. These idiosyncratic defaults are supported by the data in Table 4, which shows a concentration of tail risk (loans trading below 85) in just a few sectors: Energy, Retail, and Mining. These are the only 3 sectors that had average loan prices below 90 [5].
Table 4
2020-2021 Default Cycle
The pandemic default cycle lasted from March 2020 to March 2021. The average default rate was over 3x the long-term average default rate excluding default cycle periods, so this was a material default cycle. The cycle was defined by the severity of the sell-off in loans with average loan price plunging to 82 in March 2020. However, average loan price almost immediately recovered back into the 90s by June. This rebound was largely caused by the high volume of fiscal and monetary support provided by U.S. financial authorities.
Table 5 shows the broad distribution of tail risk (loans trading below 85) across industry sectors.
Table 5
What is the Outlook for Par-Building in 2023-2024?
2023-24 Default Cycle
In 2023, there has been a significant spike in default rates in April, May, and June driven by large defaults in healthcare services, media, and telecom. This is supported by the data in the tail risk chart in Table 6. The pace of defaults in July decreased with LCD reporting only 2 defaults for the month, both in food products. These defaults have been driven by the squeeze between high labor costs and diminished pricing power in the case of healthcare, and industry-defining secular changes in media and telecom. This concentration of defaults and price declines in just a few sectors implies that the 2023 default cycle will be more idiosyncratic (like 2019) without broad loan price declines.
However, we are forecasting the opposite. Though defaults are concentrated in just a few industries right now, larger macro concerns will start to cause defaults to creep up across industries, causing broader loan price declines.
Macro issues such as higher interest rate cost, stubborn inflation, a slowdown in economic growth and/or a short-lived recession will eventually spread the impact of leveraged loan defaults across a broader set of industries, especially to those in the B- category. As a result, we are forecasting an average loan price of 93.5 during the default cycle. This forecast is commensurate with the severity of our leveraged loan default forecast of 3.8%, which is almost 2x the long-term average during non-default periods.
Table 6
CLOs can only take advantage of par-building opportunities during their reinvestment periods, which usually last for five years after the deal is issued unless the equity investors choose to extend the window through a reset. In today’s sluggish new-issue environment, BofA estimates that 40% of U.S. CLOs will be outside of their reinvestment period by year-end 2023 [6]. This will influence the performance of CLO equity during the upcoming default cycle, as the CLO equity investors who invest in CLOs with longer reinvestment periods can benefit from par-building during default cycles. In contrast, CLO equity investors who invest in CLOs outside of their re-investment periods will see performance that is highly correlated to the credit markets and the rates of default and loss given default.
Conclusions
Par-building can be a critical support to CLO equity investors during periods of elevated defaults.
Previous default cycles offer important lessons. The 2020-2021 default cycle was probably the most unusual of the reviewed periods because the initial loan price decrease was so severe and spread fairly evenly across many industries yet par-building proved difficult due to massive fiscal and monetary support to individuals and institutions. The 2008 default cycle was the most conducive for par-building due to the severity and duration of loan price decreases.
2023-2024 cycle will have ample par-building opportunities. Based on what we know about the default cycle so far in 2023, it looks like it is impacting a narrow band of industries, driven by a handful of idiosyncratic issuers. Nevertheless, loan prices have been below the long-term average in 2023 at an average of 94.4 as of July 25 -- allowing for some par-building. We believe that loan prices will remain below the long-term average during the duration of the default cycle until mid-2024 because the bite of high nominal interest rates, stubborn inflation, and an eventual economic slowdown will impact the leveraged loan sector more broadly.
It is important for investors to keep in mind that the full effect of par-building is only available to CLOs with material time in their reinvestment windows. BofA estimates that 40% of U.S. CLOs will be outside of their reinvestment period by year-end 2023, which means that there will be many less CLOs with the ability to par-build. This could very well place a premium on deals with remaining reinvestment periods if investors believe that these deals offer par-building protections during the anticipated default cycle.
Footnotes
[1] Moody’s long-term average issuer-weighted leveraged loan default rate going back to 1996 is 3.27%.
[2] Data source: Markit leveraged loan prices and Intex CLO data. This line is based on an IRR sensitivity model of 3 actual CLO deals each with ample remaining re-investment periods. The model uses the following assumptions: long-term loan recovery rate of 60%; the sensitivity analysis is for 7 years; it uses the graph’s default rate and loan price coordinates for years 1 and 2 and then for years 3-7 it uses the historic long-term averages for defaults and loan price, excluding the periods of the default cycles; for CPR we used 14% during years 1 and 2, which is in line with historic average during the default cycles; and CPR of 20% during years 3-7 which is in line with the historic average during the periods outside of the default cycles.
[3] Except for Tables 2-6, all historic pricing data for leveraged loans comes from PitchBook LCD. All leveraged loan historic default rates are sourced from Moody’s historic issuer default data for leveraged loans.
[4] Breakdown of loan prices by industry in Tables 3 - 6 is from Markit and CLO deal data is from Intex. Data in the tables reflect the average of prices from the first and last month of the default cycle, as defined in Table 2. Data reflects the portfolio of leverage loans held by the universe of U.S. BSL CLOs.
[5] Consumer also showed prominent tail risk, but that was driven by a single large ($3 bn) default of Acosta, a consumer marketer.
[6] BofA Global Research, “Loan maturities vs CLO reinvestment end: Make Amends & Extend vs Delay & Pray.” June 9, 2023.
FORWARD-LOOKING STATEMENTS AND DISCLAIMERS
Some of the statements contained in this presentation constitute forward-looking statements within the meaning of the federal securities laws. Forward-looking statements relate to expectations, beliefs, projections, future plans and strategies, anticipated events or trends and similar expressions concerning matters that are not historical facts. In some cases, you can identify forward-looking statements by the use of forward-looking terminology such as “may,” “will,” “should,” “expects,” “intends,” “plans,” “anticipates,” “believes,” “estimates,” “predicts,” or “potential” or the negative of these words and phrases or similar words or phrases which are predictions of or indicate future events or trends and which do not relate solely to historical matters. You can also identify forward-looking statements by discussions of strategy, plans or intentions. The forward-looking statements contained in this presentation reflect our current views about future events and are subject to numerous known and unknown risks, uncertainties, assumptions and changes in circumstances, many of which are beyond our control, that may cause our actual results to differ significantly from those expressed in any forward-looking statement. Statements regarding the following subjects, among others, may be forward-looking: the use of proceeds from our public and private offerings (as the case may be); our business and investment strategy; our projected operating results; our ability to obtain financing arrangements; financing and advance rates for our target assets; our expected leverage; general volatility of the securities markets in which we invest; our expected investments; effects of hedging instruments on our target assets; rates of leasing and occupancy rates on our target assets; the degree to which our hedging strategies may or may not protect us from interest rate volatility; liquidity of our target assets; impact of changes in governmental regulations, tax law and rates, and similar matters; availability of investment opportunities; availability of qualified personnel; estimates relating to our ability to make distributions; our understanding of our competition; and market trends in our industry, interest rates, real estate values, the debt securities markets or the general economy. While forward-looking statements reflect our good faith beliefs, assumptions and expectations, they are not guarantees of future performance. Furthermore, we disclaim any obligation to publicly update or revise any forward-looking statement to reflect changes in underlying assumptions or factors, new information, data or methods, future events or other changes. This presentation contains statistics and other data that has been obtained from or compiled from information made available by third-party service providers. We have not independently verified such statistics or data. This confidential document is for informational purposes only and does not constitute an offer to sell or a solicitation of an offer to buy any securities or partnership interests described herein. Interests in Capra Credit Management, LLC (“Capra”) partnerships may not be purchased except pursuant to the partnership’s relevant subscription agreement and partnership agreement, each of which should be reviewed in its entirety prior to investment.